[AArch64] Extend tests of loads and stores of register pairs

Include instances of FP register pairs.

llvm-svn: 313638

2 files changed, 45 insertions, 2 deletions

diff --git a/llvm/test/CodeGen/AArch64/arm64-ldp.ll b/llvm/test/CodeGen/AArch64/arm64-ldp.ll
index 998ff9e895f..388f18bf801 100644
--- a/llvm/test/CodeGen/AArch64/arm64-ldp.ll
+++ b/llvm/test/CodeGen/AArch64/arm64-ldp.ll
@@ -79,6 +79,16 @@ define double @ldp_double(double* %p) nounwind {
   ret double %add
 }
 
+; CHECK-LABEL: ldp_doublex2
+; CHECK: ldp
+define <2 x double> @ldp_doublex2(<2 x double>* %p) nounwind {
+  %tmp = load <2 x double>, <2 x double>* %p, align 16
+  %add.ptr = getelementptr inbounds <2 x double>, <2 x double>* %p, i64 1
+  %tmp1 = load <2 x double>, <2 x double>* %add.ptr, align 16
+  %add = fadd <2 x double> %tmp, %tmp1
+  ret <2 x double> %add
+}
+
 ; Test the load/store optimizer---combine ldurs into a ldp, if appropriate
 define i32 @ldur_int(i32* %a) nounwind {
 ; CHECK-LABEL: ldur_int
@@ -157,7 +167,7 @@ define i64 @ldur_long(i64* %a) nounwind ssp {
 define float @ldur_float(float* %a) {
 ; CHECK-LABEL: ldur_float
 ; CHECK: ldp     [[DST1:s[0-9]+]], [[DST2:s[0-9]+]], [x0, #-8]
-; CHECK-NEXT: add     s{{[0-9]+}}, [[DST2]], [[DST1]]
+; CHECK-NEXT: fadd    s{{[0-9]+}}, [[DST2]], [[DST1]]
 ; CHECK-NEXT: ret
   %p1 = getelementptr inbounds float, float* %a, i64 -1
   %tmp1 = load float, float* %p1, align 2
@@ -170,7 +180,7 @@ define float @ldur_float(float* %a) {
 define double @ldur_double(double* %a) {
 ; CHECK-LABEL: ldur_double
 ; CHECK: ldp     [[DST1:d[0-9]+]], [[DST2:d[0-9]+]], [x0, #-16]
-; CHECK-NEXT: add     d{{[0-9]+}}, [[DST2]], [[DST1]]
+; CHECK-NEXT: fadd    d{{[0-9]+}}, [[DST2]], [[DST1]]
 ; CHECK-NEXT: ret
   %p1 = getelementptr inbounds double, double* %a, i64 -1
   %tmp1 = load double, double* %p1, align 2
@@ -180,6 +190,19 @@ define double @ldur_double(double* %a) {
   ret double %tmp3
 }
 
+define <2 x double> @ldur_doublex2(<2 x double>* %a) {
+; CHECK-LABEL: ldur_doublex2
+; CHECK: ldp     q[[DST1:[0-9]+]], q[[DST2:[0-9]+]], [x0, #-32]
+; CHECK-NEXT: fadd    v{{[0-9]+}}.2d, v[[DST2]].2d, v[[DST1]].2d
+; CHECK-NEXT: ret
+  %p1 = getelementptr inbounds <2 x double>, <2 x double>* %a, i64 -1
+  %tmp1 = load <2 x double>, <2 x double>* %p1, align 2
+  %p2 = getelementptr inbounds <2 x double>, <2 x double>* %a, i64 -2
+  %tmp2 = load <2 x double>, <2 x double>* %p2, align 2
+  %tmp3 = fadd <2 x double> %tmp1, %tmp2
+  ret <2 x double> %tmp3
+}
+
 ; Now check some boundary conditions
 define i64 @pairUpBarelyIn(i64* %a) nounwind ssp {
 ; CHECK-LABEL: pairUpBarelyIn
diff --git a/llvm/test/CodeGen/AArch64/arm64-stp.ll b/llvm/test/CodeGen/AArch64/arm64-stp.ll
index cc4591c8aec..14a54826a0d 100644
--- a/llvm/test/CodeGen/AArch64/arm64-stp.ll
+++ b/llvm/test/CodeGen/AArch64/arm64-stp.ll
@@ -36,6 +36,15 @@ define void @stp_double(double %a, double %b, double* nocapture %p) nounwind {
   ret void
 }
 
+; CHECK-LABEL: stp_doublex2
+; CHECK: stp q0, q1, [x0]
+define void @stp_doublex2(<2 x double> %a, <2 x double> %b, <2 x double>* nocapture %p) nounwind {
+  store <2 x double> %a, <2 x double>* %p, align 16
+  %add.ptr = getelementptr inbounds <2 x double>, <2 x double>* %p, i64 1
+  store <2 x double> %b, <2 x double>* %add.ptr, align 16
+  ret void
+}
+
 ; Test the load/store optimizer---combine ldurs into a ldp, if appropriate
 define void @stur_int(i32 %a, i32 %b, i32* nocapture %p) nounwind {
 ; CHECK-LABEL: stur_int
@@ -81,6 +90,17 @@ define void @stur_double(double %a, double %b, double* nocapture %p) nounwind {
   ret void
 }
 
+define void @stur_doublex2(<2 x double> %a, <2 x double> %b, <2 x double>* nocapture %p) nounwind {
+; CHECK-LABEL: stur_doublex2
+; CHECK: stp q{{[0-9]+}}, q{{[0-9]+}}, [x{{[0-9]+}}, #-32]
+; CHECK-NEXT: ret
+  %p1 = getelementptr inbounds <2 x double>, <2 x double>* %p, i32 -1
+  store <2 x double> %a, <2 x double>* %p1, align 2
+  %p2 = getelementptr inbounds <2 x double>, <2 x double>* %p, i32 -2
+  store <2 x double> %b, <2 x double>* %p2, align 2
+  ret void
+}
+
 define void @splat_v4i32(i32 %v, i32 *%p) {
 entry: